Hasil untuk "Computer applications to medicine. Medical informatics"

Wenli Guo, Xiaoli Lei, Chen Liang et al.

Mayorella marianaensis (Amoebozoa: Discosea) was the only Mayorella species isolated from deep sea (over 3000 m-depth). We firstly present a transcriptomic analysis of the non-model amoeba species collected from the Mariana Trench area. Illumina sequencing platform was used to generate data which including raw data, cleaned reads, de novo assembly, and functional annotation. After assembly, the final transcriptome consists of 57,459,544 transcripts with a mean length of 1646 and N50 length of 1170. The transcriptome has a completeness of 67.4 % as assessed by BUSCO. Functional annotation pathways related to signal transduction, transport and catabolism, and translation are the most annotated in the transcriptome.

Elizabeth A. Krupinski, Dana Abbey, Mala Muralidaran et al.

Libraries can play a valuable role in the delivery and support of telehealth services. This article is based on a panel of experts convened for the Virtual Care Symposium 2024, “From Novelty to Sustainability: How to Embed Virtual Care Into the Post-Pandemic Healthcare Delivery Template,” to discuss ways to leverage community spaces to expand telehealth access. Strategies for successful implementation and operation are provided along with insights into possible challenges based on real-world examples. Libraries can play a crucial role in bridging the digital divide by providing internet access and digital devices to the public, particularly in rural and underserved communities.

Sushant Gautam, Michael A. Riegler, Pål Halvorsen

We investigate fine-tuning Vision-Language Models (VLMs) for multi-task medical image understanding, focusing on detection, localization, and counting of findings in medical images. Our objective is to evaluate whether instruction-tuned VLMs can simultaneously improve these tasks, with the goal of enhancing diagnostic accuracy and efficiency. Using MedMultiPoints, a multimodal dataset with annotations from endoscopy (polyps and instruments) and microscopy (sperm cells), we reformulate each task into instruction-based prompts suitable for vision-language reasoning. We fine-tune Qwen2.5-VL-7B-Instruct using Low-Rank Adaptation (LoRA) across multiple task combinations. Results show that multi-task training improves robustness and accuracy. For example, it reduces the Count Mean Absolute Error (MAE) and increases Matching Accuracy in the Counting + Pointing task. However, trade-offs emerge, such as more zero-case point predictions, indicating reduced reliability in edge cases despite overall performance gains. Our study highlights the potential of adapting general-purpose VLMs to specialized medical tasks via prompt-driven fine-tuning. This approach mirrors clinical workflows, where radiologists simultaneously localize, count, and describe findings - demonstrating how VLMs can learn composite diagnostic reasoning patterns. The model produces interpretable, structured outputs, offering a promising step toward explainable and versatile medical AI. Code, model weights, and scripts will be released for reproducibility at https://github.com/simula/PointDetectCount.

Zhe Chen, Yusheng Liao, Shuyang Jiang et al.

Large language models hold promise for addressing medical challenges, such as medical diagnosis reasoning, research knowledge acquisition, clinical decision-making, and consumer health inquiry support. However, they often generate hallucinations due to limited medical knowledge. Incorporating external knowledge is therefore critical, which necessitates multi-source knowledge acquisition. We address this challenge by framing it as a source planning problem, which is to formulate context-appropriate queries tailored to the attributes of diverse sources. Existing approaches either overlook source planning or fail to achieve it effectively due to misalignment between the model's expectation of the sources and their actual content. To bridge this gap, we present MedOmniKB, a repository comprising multigenre and multi-structured medical knowledge sources. Leveraging these sources, we propose the Source Planning Optimisation method, which enhances multi-source utilisation. Our approach involves enabling an expert model to explore and evaluate potential plans while training a smaller model to learn source alignment. Experimental results demonstrate that our method substantially improves multi-source planning performance, enabling the optimised small model to achieve state-of-the-art results in leveraging diverse medical knowledge sources.

Siyi Xun, Yue Sun, Jingkun Chen et al.

Rapid advances in medical imaging technology underscore the critical need for precise and automated image quality assessment (IQA) to ensure diagnostic accuracy. Existing medical IQA methods, however, struggle to generalize across diverse modalities and clinical scenarios. In response, we introduce MedIQA, the first comprehensive foundation model for medical IQA, designed to handle variability in image dimensions, modalities, anatomical regions, and types. We developed a large-scale multi-modality dataset with plentiful manually annotated quality scores to support this. Our model integrates a salient slice assessment module to focus on diagnostically relevant regions feature retrieval and employs an automatic prompt strategy that aligns upstream physical parameter pre-training with downstream expert annotation fine-tuning. Extensive experiments demonstrate that MedIQA significantly outperforms baselines in multiple downstream tasks, establishing a scalable framework for medical IQA and advancing diagnostic workflows and clinical decision-making.

Luisa Gallée, Yiheng Xiong, Meinrad Beer et al.

Densely annotated medical image datasets that capture not only diagnostic labels but also the underlying reasoning behind these diagnoses are scarce. Such reasoning-related annotations are essential for developing and evaluating explainable AI (xAI) models that reason similarly to radiologists: making correct predictions for the right reasons. To address this gap, we introduce FunnyNodules, a fully parameterized synthetic dataset designed for systematic analysis of attribute-based reasoning in medical AI models. The dataset generates abstract, lung nodule-like shapes with controllable visual attributes such as roundness, margin sharpness, and spiculation. The target class is derived from a predefined attribute combination, allowing full control over the decision rule that links attributes to the diagnostic class. We demonstrate how FunnyNodules can be used in model-agnostic evaluations to assess whether models learn correct attribute-target relations, to interpret over- or underperformance in attribute prediction, and to analyze attention alignment with attribute-specific regions of interest. The framework is fully customizable, supporting variations in dataset complexity, target definitions, class balance, and beyond. With complete ground truth information, FunnyNodules provides a versatile foundation for developing, benchmarking, and conducting in-depth analyses of explainable AI methods in medical image analysis.

Luke Flanagan, Maija Poikela

Stigma is a considerable barrier to accessing healthcare services for stigmatized communities. Digital health technologies support interventions that either aim at reducing stigma (or its perception), or alternatively, providing healthcare in a stigma-accommodating manner. Studies report mixed results, with positive evidence that stigma-reducing interventions have an impact, but the few large-scale studies show weaker effects. The growing field of digital health presents an opportunity for novel stigma-reduction methodologies at scale, as well as to implement novel methods to circumvent stigma-related barriers, improving direct access to healthcare services. We review the barriers to healthcare as a consequence of stigma; examine a sample of recent interventions; and propose future work to enhance healthcare access.

Gurucharan Marthi Krishna Kumar, Aman Chadha, Janine Mendola et al.

Large Language Models (LLMs), known for their versatility in textual data, are increasingly being explored for their potential to enhance medical image segmentation, a crucial task for accurate diagnostic imaging. This study explores enhancing Vision Transformers (ViTs) for medical image segmentation by integrating pre-trained LLM transformer blocks. Our approach, which incorporates a frozen LLM transformer block into the encoder of a ViT-based model, leads to substantial improvements in segmentation performance across various medical imaging modalities. We propose a Hybrid Attention Mechanism that combines global and local feature learning with a Multi-Scale Fusion Block for aggregating features across different scales. The enhanced model shows significant performance gains, including an average Dice score increase from 0.74 to 0.79 and improvements in accuracy, precision, and the Jaccard Index. These results demonstrate the effectiveness of LLM-based transformers in refining medical image segmentation, highlighting their potential to significantly boost model accuracy and robustness. The source code and our implementation are available at: https://github.com/AS-Lab/Marthi-et-al-2025-MedVisionLlama-Pre-Trained-LLM-Layers-to-Enhance-Medical-Image-Segmentation

Jamil Zaghir, Marco Naguib, Mina Bjelogrlic et al.

Prompt engineering is crucial for harnessing the potential of large language models (LLMs), especially in the medical domain where specialized terminology and phrasing is used. However, the efficacy of prompt engineering in the medical domain remains to be explored. In this work, 114 recent studies (2022-2024) applying prompt engineering in medicine, covering prompt learning (PL), prompt tuning (PT), and prompt design (PD) are reviewed. PD is the most prevalent (78 articles). In 12 papers, PD, PL, and PT terms were used interchangeably. ChatGPT is the most commonly used LLM, with seven papers using it for processing sensitive clinical data. Chain-of-Thought emerges as the most common prompt engineering technique. While PL and PT articles typically provide a baseline for evaluating prompt-based approaches, 64% of PD studies lack non-prompt-related baselines. We provide tables and figures summarizing existing work, and reporting recommendations to guide future research contributions.

Paul Knowles, Bart Gajderowicz, Keith Dugas

The Core Data Ontology (CDO) and the Informatics Domain Model represent a transformative approach to computational systems, shifting from traditional node-centric designs to a data-centric paradigm. This paper introduces a framework where data is categorized into four modalities: objects, events, concepts, and actions. This quadrimodal structure enhances data security, semantic interoperability, and scalability across distributed data ecosystems. The CDO offers a comprehensive ontology that supports AI development, role-based access control, and multimodal data management. By focusing on the intrinsic value of data, the Informatics Domain Model redefines system architectures to prioritize data security, provenance, and auditability, addressing vulnerabilities in current models. The paper outlines the methodology for developing the CDO, explores its practical applications in fields such as AI, robotics, and legal compliance, and discusses future directions for scalable, decentralized, and interoperable data ecosystems.

M. Laroussi, Sander Bekeschus, M. Keidar et al.

Plasma, the fourth and most pervasive state of matter in the visible universe, is a fascinating medium that is connected to the beginning of our universe itself. Man-made plasmas are at the core of many technological advances that include the fabrication of semiconductor devices, which enabled the modern computer and communication revolutions. The introduction of low temperature, atmospheric pressure plasmas to the biomedical field has ushered a new revolution in the healthcare arena that promises to introduce plasma-based therapies to combat some thorny and long-standing medical challenges. This article presents an overview of where research is at today and discusses innovative concepts and approaches to overcome present challenges and take the field to the next level. It is written by a team of experts who took an in-depth look at the various applications of plasma in hygiene, decontamination, and medicine, made critical analysis, and proposed ideas and concepts that should help the research community focus their efforts on clear and practical steps necessary to keep the field advancing for decades to come.

Fabiana Sofia Ricci, Alain Boldini, Xinda Ma et al.

Visual impairment represents a significant health and economic burden affecting 596 million globally. The incidence of visual impairment is expected to double by 2050 as our population ages. Independent navigation is challenging for persons with visual impairment, as they often rely on non-visual sensory signals to find the optimal route. In this context, electronic travel aids are promising solutions that can be used for obstacle detection and/or route guidance. However, electronic travel aids have limitations such as low uptake and limited training that restrict their widespread use. Here, we present a virtual reality platform for testing, refining, and training with electronic travel aids. We demonstrate the viability on an electronic travel aid developed in-house, consist of a wearable haptic feedback device. We designed an experiment in which participants donned the electronic travel aid and performed a virtual task while experiencing a simulation of three different visual impairments: age-related macular degeneration, diabetic retinopathy, and glaucoma. Our experiments indicate that our electronic travel aid significantly improves the completion time for all the three visual impairments and reduces the number of collisions for diabetic retinopathy and glaucoma. Overall, the combination of virtual reality and electronic travel aid may have a beneficial role on mobility rehabilitation of persons with visual impairment, by allowing early-phase testing of electronic travel aid prototypes in safe, realistic, and controllable settings. Author summary Testing electronic travel aids under development is an outstanding research area, due to the rapid growth in the number of people with visual impairment. For decades, different technologies have been employed to improve the mobility of persons with visual impairment, but suitable and easy-to-use solutions have not yet been established. In this study, we propose the use of virtual reality for early-phase testing of electronic travel aids in safe, realistic, and controllable settings. We demonstrate the approach using an haptic feedback device developed in-house in the form of a belt. Our device can be simply wear by a user, providing free hands and real-time operation. The approach offers the combined possibility of designing highly realistic, urban environments and of simulating different forms of visual impairment on healthy subjects. Our integrated wearable electronic travel aid/virtual reality system establishes a novel assistive framework to mitigate the consequences of visual impairment. We envision this framework could improve training, reduce rehabilitation, and abate societal costs, while creating an engaging and compelling experience for persons with visual impairment.

Saikat Roy, Gregor Koehler, Michael Baumgartner et al.

Owing to success in the data-rich domain of natural images, Transformers have recently become popular in medical image segmentation. However, the pairing of Transformers with convolutional blocks in varying architectural permutations leaves their relative effectiveness to open interpretation. We introduce Transformer Ablations that replace the Transformer blocks with plain linear operators to quantify this effectiveness. With experiments on 8 models on 2 medical image segmentation tasks, we explore -- 1) the replaceable nature of Transformer-learnt representations, 2) Transformer capacity alone cannot prevent representational replaceability and works in tandem with effective design, 3) The mere existence of explicit feature hierarchies in transformer blocks is more beneficial than accompanying self-attention modules, 4) Major spatial downsampling before Transformer modules should be used with caution.

Yuhao Huang, Xin Yang, Lian Liu et al.

The Segment Anything Model (SAM) is the first foundation model for general image segmentation. It has achieved impressive results on various natural image segmentation tasks. However, medical image segmentation (MIS) is more challenging because of the complex modalities, fine anatomical structures, uncertain and complex object boundaries, and wide-range object scales. To fully validate SAM's performance on medical data, we collected and sorted 53 open-source datasets and built a large medical segmentation dataset with 18 modalities, 84 objects, 125 object-modality paired targets, 1050K 2D images, and 6033K masks. We comprehensively analyzed different models and strategies on the so-called COSMOS 1050K dataset. Our findings mainly include the following: 1) SAM showed remarkable performance in some specific objects but was unstable, imperfect, or even totally failed in other situations. 2) SAM with the large ViT-H showed better overall performance than that with the small ViT-B. 3) SAM performed better with manual hints, especially box, than the Everything mode. 4) SAM could help human annotation with high labeling quality and less time. 5) SAM was sensitive to the randomness in the center point and tight box prompts, and may suffer from a serious performance drop. 6) SAM performed better than interactive methods with one or a few points, but will be outpaced as the number of points increases. 7) SAM's performance correlated to different factors, including boundary complexity, intensity differences, etc. 8) Finetuning the SAM on specific medical tasks could improve its average DICE performance by 4.39% and 6.68% for ViT-B and ViT-H, respectively. We hope that this comprehensive report can help researchers explore the potential of SAM applications in MIS, and guide how to appropriately use and develop SAM.

Gabriel F. C. de Queiroz, José F. de Rezende, Valmir C. Barbosa

Multi-access Edge Computing (MEC) is an enabling technology to leverage new network applications, such as virtual/augmented reality, by providing faster task processing at the network edge. This is done by deploying servers closer to the end users to run the network applications. These applications are often intensive in terms of task processing, memory usage, and communication; thus mobile devices may take a long time or even not be able to run them efficiently. By transferring (offloading) the execution of these applications to the servers at the network edge, it is possible to achieve a lower completion time (makespan) and meet application requirements. However, offloading multiple entire applications to the edge server can overwhelm its hardware and communication channel, as well as underutilize the mobile devices' hardware. In this paper, network applications are modeled as Directed Acyclic Graphs (DAGs) and partitioned into tasks, and only part of these tasks are offloaded to the edge server. This is the DAG application partitioning and offloading problem, which is known to be NP-hard. To approximate its solution, this paper proposes the FlexDO algorithm. FlexDO combines a greedy phase with a permutation phase to find a set of offloading decisions, and then chooses the one that achieves the shortest makespan. FlexDO is compared with a proposal from the literature and two baseline decisions, considering realistic DAG applications extracted from the Alibaba Cluster Trace Program. Results show that FlexDO is consistently only 3.9% to 8.9% above the optimal makespan in all test scenarios, which include different levels of CPU availability, a multi-user case, and different communication channel transmission rates. FlexDO outperforms both baseline solutions by a wide margin, and is three times closer to the optimal makespan than its competitor.

Mario Mascalchi

MRI shows the three archetypal patterns of CNS volume loss underlying progressive ataxias in vivo, namely spinal atrophy (SA), cortical cerebellar atrophy (CCA) and olivopontocerebellar atrophy (OPCA). The MRI-based CNS atrophy pattern was reviewed in 128 progressive ataxias. A CNS atrophy pattern was identified in 91 conditions: SA in Friedreich’s ataxia, CCA in 5 acquired and 72 (24 dominant, 47 recessive,1 X-linked) inherited ataxias, OPCA in Multi-System Atrophy and 12 (9 dominant, 2 recessive,1 X-linked) inherited ataxias. The MRI-based CNS atrophy pattern may be useful for genetic assessment, identification of shared cellular targets, repurposing therapies or the enlargement of drug indications in progressive ataxias.

Louise G. Køhler, Bastian Huwer, José Martín Pujolar et al.

This data article includes a qualitative and quantitative description of the gelatinous macrozooplankton community of the North Sea during January-February 2021. Sampling was conducted during the 1st quarter International Bottom Trawl Survey (IBTS) on board the Danish R/V DANA (DTU Aqua Denmark) and the Swedish R/V Svea (SLU Sweden), as part of the ichthyoplankton investigation during night-time. A total of 147 stations were investigated in the western, central and eastern North Sea as well as the Skagerrak and Kattegat. Sampling was conducted with a 13 m long Midwater Ring Net (MIK net, Ø 2 m, mesh size 1.6 mm, cod end with smaller mesh size of 500 µm), equipped with a flow meter. The MIK net was deployed in double oblique hauls from the surface to c. 5 m above the sea floor [1,2]. Samples were visually analysed unpreserved on a light table and/or with a stereomicroscope or magnifying lamp within 2 hours after catch. A total of 13,510 individuals were counted/sized. Twelve gelatinous macrozooplankton species or genera were encountered, namely the hydrozoan Aequorea vitrina, Aglantha digitale, Clytia spp., Leuckartiara octona, Tima bairdii, Muggiaea atlantica; the scyphozoans Cyanea capillata and Cyanea lamarckii and the ctenophores Beroe spp., Bolinopsis infundibulum, Mnemiopsis leidyi, Pleurobrachia pileus. Abundance data are presented on a volume specific (m−3) and area specific (m−2) basis. Size data have been used to estimate wet weights based on published length-weight regressions (see Table 1). For the groups i) hydrozoan jellyfish, ii) scyphozoan jellyfish, iii) ctenophores, as well as iv) grouped gelatinous macrozooplankton, spatial weight specific distribution patterns are presented. This unpublished dataset contributes baseline information about the gelatinous macrozooplankton diversity and its specific distribution patterns in the extended North Sea area during winter (January-February) 2021. These data can be an important contribution to address global change impacts on marine systems, especially considering gelatinous macrozooplankton abundance changes in relation to anthropogenic stressors.

Van C Willis, Kelly Jean Thomas Craig, Yalda Jabbarpour et al.

BackgroundDisease prevention is a central aspect of primary care practice and is comprised of primary (eg, vaccinations), secondary (eg, screenings), tertiary (eg, chronic condition monitoring), and quaternary (eg, prevention of overmedicalization) levels. Despite rapid digital transformation of primary care practices, digital health interventions (DHIs) in preventive care have yet to be systematically evaluated. ObjectiveThis review aimed to identify and describe the scope and use of current DHIs for preventive care in primary care settings. MethodsA scoping review to identify literature published from 2014 to 2020 was conducted across multiple databases using keywords and Medical Subject Headings terms covering primary care professionals, prevention and care management, and digital health. A subgroup analysis identified relevant studies conducted in US primary care settings, excluding DHIs that use the electronic health record (EHR) as a retrospective data capture tool. Technology descriptions, outcomes (eg, health care performance and implementation science), and study quality as per Oxford levels of evidence were abstracted. ResultsThe search yielded 5274 citations, of which 1060 full-text articles were identified. Following a subgroup analysis, 241 articles met the inclusion criteria. Studies primarily examined DHIs among health information technologies, including EHRs (166/241, 68.9%), clinical decision support (88/241, 36.5%), telehealth (88/241, 36.5%), and multiple technologies (154/241, 63.9%). DHIs were predominantly used for tertiary prevention (131/241, 54.4%). Of the core primary care functions, comprehensiveness was addressed most frequently (213/241, 88.4%). DHI users were providers (205/241, 85.1%), patients (111/241, 46.1%), or multiple types (89/241, 36.9%). Reported outcomes were primarily clinical (179/241, 70.1%), and statistically significant improvements were common (192/241, 79.7%). Results were summarized across the following 5 topics for the most novel/distinct DHIs: population-centered, patient-centered, care access expansion, panel-centered (dashboarding), and application-driven DHIs. The quality of the included studies was moderate to low. ConclusionsPreventive DHIs in primary care settings demonstrated meaningful improvements in both clinical and nonclinical outcomes, and across user types; however, adoption and implementation in the US were limited primarily to EHR platforms, and users were mainly clinicians receiving alerts regarding care management for their patients. Evaluations of negative results, effects on health disparities, and many other gaps remain to be explored.

Konstantin Butenko, Ningfei Li, Clemens Neudorfer et al.

Background: Deep brain stimulation (DBS) is an established therapy for patients with Parkinson’s disease. In silico computer models for DBS hold the potential to inform a selection of stimulation parameters. In recent years, the focus has shifted towards DBS-induced firing in myelinated axons, deemed particularly relevant for the external modulation of neural activity.Objective: The aim of this project was to investigate correlations between patient-specific pathway activation profiles and clinical motor improvement.Methods: We used the concept of pathway activation modeling, which incorporates advanced volume conductor models and anatomically authentic fiber trajectories to estimate DBS-induced action potential initiation in anatomically plausible pathways that traverse in close proximity to targeted nuclei. We applied the method on two retrospective datasets of DBS patients, whose clinical improvement had been evaluated according to the motor part of the Unified Parkinson’s Disease Rating Scale. Based on differences in outcome and activation levels for intrapatient DBS protocols in a training cohort, we derived a pathway activation profile that theoretically induces a complete alleviation of symptoms described by UPDRS-III. The profile was further enhanced by analyzing the importance of matching activation levels for individual pathways.Results: The obtained profile emphasized the importance of activation in pathways descending from the motor-relevant cortical regions as well as the pallidothalamic pathways. The degree of similarity of patient-specific profiles to the optimal profile significantly correlated with clinical motor improvement in a test cohort.Conclusion: Pathway activation modeling has a translational utility in the context of motor symptom alleviation in Parkinson’s patients treated with DBS.

S. Srinivasan, M. Babu, P.S. Shabi Shimny et al.

The commodity futures market plays a major role in reducing the price risk for the participants. Unfortunately, the farmers’ participation in the futures market particularly from the Tamil Nadu region is very less. A survey was conducted using the interview method to identify the information sources used by farmers for taking pricing decisions, the awareness and perception of farmers towards the futures market, and its effect on preferred marketing alternatives. The data cleaning process was done using content validity, confirmatory factor analysis, and reliability test using Cronbach's alpha, and the assumptions of normality and multicollinearity were examined. The data will be of potential use to researchers who wish to explore farmers’ behavior towards hedging in the commodity futures market.